.yaml files in the models folder

BICLab / EMS-YOLO

Offical implementation of "Deep Directly-Trained Spiking Neural Networks for Object Detection" (ICCV2023)

https://arxiv.org/abs/2307.11411

GNU General Public License v3.0

139 stars 12 forks source link

.yaml files in the models folder #8

Open leozoro opened 7 months ago

leozoro commented 7 months ago

I'm confused about the .yaml files in the models folder. I don't know what model does each .yaml file correspond to. Could anyone help me? Thanks.

JDYG commented 5 months ago

I have the same problem. As the author state in the README.md, they use the ResNet34 model, which I guess should correspond to the resnet34.yaml. However, when I check the content in the resnet34.yaml, I found that only BasicBlock_2 is used in the backbone, and the MS_block isn't used in the network. I think this is not align with what they presented in the paper.

Hope someone could provide some detials.

JDYG commented 5 months ago

I also noticed a discrepancy between Concat_res2 and the 'EMS_block2' described in the paper. In the paper, the Maxpool operation is applied first followed by the LCB block. However, in the Concat_res2 code, the LCB block is done first, and then the Maxpool operation is performed after concatenation.